Message boards :
News :
SETI@home Version 7 has been released
Message board moderation
Previous · 1 . . . 12 · 13 · 14 · 15 · 16 · 17 · 18 . . . 19 · Next
Author | Message |
---|---|
CPY Send message Joined: 26 Aug 05 Posts: 3 Credit: 232,622 RAC: 0 |
So... what changed? I'm waiting for that post :) I'm very curious about it! I see i'm not alone who is curious :D |
kittyman Send message Joined: 9 Jul 00 Posts: 51484 Credit: 1,018,363,574 RAC: 1,004 |
I have sent a message to Eric telling him that we are all very interested in his explanations about v7. I know he was really, really stretched thin during the first few days of the conversion. But I don't know what else he may have had on his plate this week or if he is in the lab. I know he intended to post about it before this. Please give him time...... "Time is simply the mechanism that keeps everything from happening all at once." |
Alex2 Send message Joined: 2 Nov 12 Posts: 24 Credit: 11,798,226 RAC: 0 |
I'd like to get a hint of why v7 and Lunatics v7 are 30% slower on my Cuda5? |
Igogo Send message Joined: 18 Dec 04 Posts: 125 Credit: 65,303,299 RAC: 44 |
I'd like to get a hint of why v7 and Lunatics v7 are 30% slower on my Cuda5? Confirm this. I don't know precition value 30% or other but slower |
Gatekeeper Send message Joined: 14 Jul 04 Posts: 887 Credit: 176,479,616 RAC: 0 |
I'd like to get a hint of why v7 and Lunatics v7 are 30% slower on my Cuda5? Mostly because of additional processing with v7 vs v6. And 30% is (IMO) below average. Even "shorties" on my 5xx GPU's are running double or more the previous times. |
Richard Haselgrove Send message Joined: 4 Jul 99 Posts: 14680 Credit: 200,643,578 RAC: 874 |
I'd like to get a hint of why v7 and Lunatics v7 are 30% slower on my Cuda5? The extra amount of processing required is pretty constant across all ARs. So it actually shows up most (it's a higher %age increase) for shorties - it's less noticeable at lower ARs. |
Bill Butler Send message Joined: 26 Aug 03 Posts: 101 Credit: 4,270,697 RAC: 0 |
But I don't know what else he may have had on his plate this week or if he is in the lab. There is the 222nd meeting of the American Astronomical Society this week in Indianapolis. See an article mentioning it here "It is often darkest just before it turns completely black." |
Von Send message Joined: 11 Mar 13 Posts: 2 Credit: 84,641 RAC: 0 |
I getting a notice to upgrade the gpu. I also keep getting computation error on everything. Anyone know whats going on |
kittyman Send message Joined: 9 Jul 00 Posts: 51484 Credit: 1,018,363,574 RAC: 1,004 |
I getting a notice to upgrade the gpu. I also keep getting computation error on everything. Anyone know whats going on A notice to upgrade the GPU? From where or what? I don't think Boinc has any such messages. "Time is simply the mechanism that keeps everything from happening all at once." |
Grant (SSSF) Send message Joined: 19 Aug 99 Posts: 13861 Credit: 208,696,464 RAC: 304 |
Old video hardware, older driver. Error message on the errored WUs- too many exits. Grant Darwin NT |
Martin Berndt Send message Joined: 20 May 99 Posts: 4 Credit: 2,647,202 RAC: 0 |
I'd like to get a hint of why v7 and Lunatics v7 are 30% slower on my Cuda5? Hmmm... Even my Geforce GTX 650 (2 GB GDDR5) is now 30% SLOWER!!! :( |
rob smith Send message Joined: 7 Mar 03 Posts: 22556 Credit: 416,307,556 RAC: 380 |
If you care to look through "number crunching" you will find an explanation (several threads) - In short, there are some new "features" added which add an overhead to the task duration, the processing time for these far exceeds the performance gain from the core task. Bob Smith Member of Seti PIPPS (Pluto is a Planet Protest Society) Somewhere in the (un)known Universe? |
Martin Berndt Send message Joined: 20 May 99 Posts: 4 Credit: 2,647,202 RAC: 0 |
If you care to look through "number crunching" you will find an explanation (several threads) - In short, there are some new "features" added which add an overhead to the task duration, the processing time for these far exceeds the performance gain from the core task. Well... I can also tell it much easier: SHIT CODE! :p Please give us back v6! |
Raistmer Send message Joined: 16 Jun 01 Posts: 6325 Credit: 106,370,077 RAC: 121 |
If you care to look through "number crunching" you will find an explanation (several threads) - In short, there are some new "features" added which add an overhead to the task duration, the processing time for these far exceeds the performance gain from the core task. I would say some actively refuse to read and understand what other wrote. NEW SEARCH ADDED ! Comprehend ? SETI apps news We're not gonna fight them. We're gonna transcend them. |
Martin Berndt Send message Joined: 20 May 99 Posts: 4 Credit: 2,647,202 RAC: 0 |
Wasting GPU power for nothing --> disabled Maybe some months later again if they fix the worst bugs! |
Claggy Send message Joined: 5 Jul 99 Posts: 4654 Credit: 47,537,079 RAC: 4 |
Wasting GPU power for nothing --> disabled You're not wasting GPU power, you're producing good science with your GPU, you haven't errored any GPU tasks out, so you haven't encountered any Bugs: All tasks for computer 6504517 Application details for host 6504517 Just because the the autocorrection code isn't ultra fast doesn't mean it's Buggy, it's just New, expect over the next couple of years that the autocorrection code, as well as a lot of the other code will get further optimised as the volunteer developers experiment, and if it wasn't for the volunteer developers, we would only have v7 CPU apps and slower CPU apps at that. When you're devised a better method of doing the autocorrection search please let us know, then it can be fed back into the source. Claggy |
Thomas Send message Joined: 9 Dec 11 Posts: 1499 Credit: 1,345,576 RAC: 0 |
Wasting GPU power for nothing --> disabled Well said Claggy ! +100 !!! I love >> You're not wasting GPU power, you're producing good science with your GPU For my part, I spend at NVIDIA next week ! Finally ! :) |
Martin Berndt Send message Joined: 20 May 99 Posts: 4 Credit: 2,647,202 RAC: 0 |
Well... I would also help if I could, so give me your source! :p (Yes, I'm assembler programmer already 30 years. And bugs not only create errors.) |
Claggy Send message Joined: 5 Jul 99 Posts: 4654 Credit: 47,537,079 RAC: 4 |
It's on the front page, Porting and optimizing SETI@home You want the XBranch for the Cuda app: branches/sah_v7_opt @ 1862 Claggy |
Gundolf Jahn Send message Joined: 19 Sep 00 Posts: 3184 Credit: 446,358 RAC: 0 |
|
©2024 University of California
SETI@home and Astropulse are funded by grants from the National Science Foundation, NASA, and donations from SETI@home volunteers. AstroPulse is funded in part by the NSF through grant AST-0307956.